What is so special with the powerset operation? (Q1882625)

The author investigates positive operators defined as follows. Given a second-order formula \(\phi(v,c_1,\dots,c_n,S)\) of the language of set theory, where \(S\) is a class variable, \(v\) is a set variable and \((c_1,\dots,c_n)\) is a tuple of set parameters, the operator \(\Gamma\) induced by \(\phi\) with parameters \(c_1,\dots,c_n\) is a mapping from classes to classes defined by \(\Gamma(X)=\{ x:\phi(x,c_1,\dots,c_n,X)\}\). (Such a \(\Gamma\) is a set-operator if it sends sets to sets.) \(\Gamma\) is a positive operator if the formula \(\phi\) is constructed from formulas not containing \(S\) and atomic formulas \(u\in S\) using only the logical operations \(\wedge\), \(\vee\), \(\exists\) and \(\forall\). The powerset operator \(\mathcal P\) is a positive set-operator which raises the cardinality of its arguments, i.e., \(| {\mathcal P}(x)| >| x| \). Operators with these properties are called \(\mathcal P\)-like. One of the two main results of this article is a normal-form theorem which states that every positive operator is a combination of the identity operator, the powerset operator and almost constant operators (those whose ranges are set-size families of classes). Here the allowable combinations include recursive applications of composition and finitary and uniformly infinite unions and intersections. A positive operator which is a combination of the identity operator and almost constant operators is thus naturally called \(\mathcal P\)-independent. The author conjectures that every \(\mathcal P\)-independent positive operator is not \(\mathcal P\)-like, but is only able to verify it for \(\mathcal P\)-independent bounded positive operators (in whose definitions only bounded quantifications are used).